trellis-mac
AI
image-to-3d generation
SHOW HN2026
How HN took it
7 positive, 6 neutral or asking, 0 mixed, 4 critical, out of 17 comments about the project. Thread score +0.14.
Idea: 4 for, 1 againstExecution: 1 for, 2 againstDocs: 0 for, 1 against
The thread, comment by comment
Each row is a top-level comment and its replies, in HN's order. Select a cell to open that comment on HN.
positive neutral or mixed critical about another commenter or something else the maker
What HN said
What people liked
good porting work replacing CUDA-only kernels with pure PyTorch
From 3 comments
This is the kind of porting work I always hope for when I see a CUDA-only release. Have you thought about publishing the gather-scatter sparse 3D convolution and SDPA attention…
sebakubisz on HNMost 'runs on Mac' ports are a wrapper around a cloud call or a quantized shell of the original model. Going after the CUDA-specific kernels with pure-PyTorch alternatives is the…
sergiopreira on HNsolid port. the sdpa swap for sparse attention — did you notice a meaningful quality difference, or is it basically equivalent to the cuda version? curious if the pure-pytorch…
paolatauru on HN
Objections
much slower than H100, about 3.5 minutes per mesh
From 2 comments
I thought it was cool and then I found the open issue mentioned above, that convinced me its def something more people want. It IS significantly slower, about 3.5 minutes on my…
shivampkumar on HNi was able to get it in 3.5 mins from a single image on my 24gb m4 pro macbook I'm still working on this to try to replicate nvdiffrast better. Found an open source port, might…
shivampkumar on HN
TRELLIS.2 output quality is weak compared with other image-to-3D models
From 1 comment
Nice work. Although this model is not very good, I tried a lot of different image-to-3d models, the one from meshy.ai is the best, trellis is in the useless tier, really hope…
gondar on HN
high memory use, about 15GB at peak, rules out 8GB Macs
From 1 comment
The model needed about 15GB at peak during generation - the 4B model loads multiple sub-models (1.3B each for shape and texture flow). 8GB won't be enough, but both 24GB and 32GB…
shivampkumar on HN
Requests
add examples and images to the landing page
From 1 comment
So much effort, but no examples in the landing page.
kennyloginz on HN
Metal shader implementations for speed
From 1 comment
Great. Potentially can go much faster rewriting it in terms of Metal shaders.
antirez on HN
publish the sparse conv and attention swaps as a reusable toolkit
From 1 comment
This is the kind of porting work I always hope for when I see a CUDA-only release. Have you thought about publishing the gather-scatter sparse 3D convolution and SDPA attention…
sebakubisz on HN
Questions
how much RAM is needed
From 1 comment
How much RAM does this use? Only sitting on 8 GB right now, I'm trying to figure out if I should buy 24 GB when it's time for a replacement or spring for 32.
post-it on HN
can sub-models load in series to fit 16GB
From 1 comment
Thanks! Could it conceivably load the sub-models in series rather than parallel? 8 still won't be enough but I wonder if those with 16 could eke something out.
post-it on HN
does it support multi-view input
From 1 comment
Does it support multi-view input?
drbscl on HN
Alternatives named
Meshy.ai, a paid hosted image-to-3D service
From 2 comments
Nice work. Although this model is not very good, I tried a lot of different image-to-3d models, the one from meshy.ai is the best, trellis is in the useless tier, really hope…
gondar on HNMeshy is indeed great but I am terminally put off by their alltogether terrible, sleazy, gamified, opaque web UI. It's like aliexpress and a lootbox game had a baby that's into…
isoprophlex on HN
Apple ml-sharp, which creates Gaussian splats rather than meshes
From 2 comments
I wonder how these outputs compare between Apple Sharp https://github.com/apple/ml-sharp No matter what it is cool seeing so much them work on different devices
pcoyne on HNThey’re completely different projects. ML Sharp is designed to create a 3D Gaussian representation of a depicted scene and output a splat file. Trellis generates model files (GLB…
vunderba on HN